JP2013249506A - Symbol identification method and symbol identification device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To recognize a symbol for individual identification having measured temperature different from surface temperature of an object by using a two-dimentional temperature data measured with an infrared camera.SOLUTION: A symbol 3 for individual identification drawn with a heat-resistant paint having different emissivity, on a shell surface of a ladle 2, is measured with an infrared camera 4, a two-dimentional temperature data acquisition part 11 extracts temperature data of a symbol identification region including the symbol 3, a filtering part 12 removes noise, and a binary coded processing part 13 binarizes the noise-removed temperature data of the symbol identification region using a prescribed temperature threshold value and converts the binarized temperature data into an input image. A feature-conformity determination part 15 collates an input image with reference images of the respective symbols prepared in advance in a reference image holding part 14, and determines the symbol of the reference image having high degree of conformity as a symbol identification result.

Description

  The present invention relates to a symbol identification method and a symbol identification device for identifying a symbol from temperature data obtained by measuring a target area with an infrared camera.

  The ladle that stores molten steel melted in a converter or electric furnace and conveys it to the next process deteriorates due to wear due to temperature change, and in the worst case, the molten steel may leak. Therefore, conventionally, a steel leakage prevention method has been proposed in which the iron skin temperature on the ladle surface is measured with an infrared camera, the possibility of leakage is predicted based on the measured temperature, and leakage is prevented in advance (for example, a patent) Reference 1).

  On the other hand, the individual identification of the ladle has been such that the operator visually confirms the symbol drawn on the surface of the ladle or recognizes characters from the image captured by the optical camera. In addition, as a character recognition method, although a use differs with the symbol of a ladle and a vehicle license plate, there exists a method of character-recognizing the license plate image imaged with the infrared camera (for example, refer patent document 2).

JP 2009-19249 A JP 2009-245385 A

  In the above-mentioned Patent Document 1, only the measured two-dimensional temperature data is used for leak prediction, and there is a problem that a dedicated device such as an optical camera is required for individual identification.

  On the other hand, in the above-mentioned patent document 2, the recognition target is a vehicle license plate, and the image picked up by the infrared camera is mainly used instead at night when it is difficult to pick up with the optical camera. Therefore, it is assumed that character recognition of an infrared camera image is performed by the same method as character recognition of a conventional optical camera image, and character recognition based on temperature data is not performed.

  The present invention has been made to solve the above-described problems, and uses two-dimensional temperature data measured by an infrared camera to recognize an individual identification symbol whose temperature is different from the surface temperature of an object. An object is to provide an identification method and a symbol identification device.

  The symbol identification method according to the present invention includes a two-dimensional temperature data acquisition step of acquiring two-dimensional temperature data obtained by measuring an object having an individual identification symbol attached to a surface using materials having different emissivities with an infrared camera. A binarization processing step for binarizing the two-dimensional temperature data acquired in the two-dimensional temperature data acquisition step with a predetermined temperature threshold to generate an input image, and an infrared camera for all types of symbols for individual identification A reference image measured and binarized in advance is prepared in advance, and each of the reference images is collated with the input image generated in the binarization processing step, and the symbol of the reference image having the highest degree of matching is input to the input image. And a feature matching determination step for recognizing it as a symbol.

  A symbol identification device according to the present invention includes a two-dimensional temperature data acquisition unit that acquires two-dimensional temperature data obtained by measuring an object having an individual identification symbol attached to a surface using materials having different emissivities with an infrared camera. A binarization processing unit that binarizes the two-dimensional temperature data acquired by the two-dimensional temperature data acquisition unit with a predetermined temperature threshold and generates an input image, and an infrared camera for all types of symbols for individual identification The reference image holding unit holding the reference image measured and binarized and the reference image held by the reference image holding unit are compared with the input image generated by the binarization processing unit. A feature matching determination unit that recognizes a symbol of a reference image having a high matching degree as a symbol of the input image.

  According to the present invention, since the individual identification symbol having an emissivity different from that of the surface of the object, that is, the measured temperature is recognized from the two-dimensional temperature data measured by the infrared camera, the temperature management of the object is performed. Infrared cameras can be used for individual identification, and a low-cost and high added-value symbol identification method and symbol identification device can be provided.

It is a figure which shows the structure of the symbol identification system containing the symbol identification device which concerns on Embodiment 1 of this invention. It is a figure which shows an example of the two-dimensional temperature data which the symbol identification device which concerns on Embodiment 1 uses. It is a figure which shows the example of the reference | standard image which the reference | standard image holding | maintenance part hold | maintains. It is a figure which shows the example in which a feature matching determination part converts 2D data into a 1-dimensional data sequence. FIGS. 5A and 5B are diagrams illustrating processing of a feature matching determination unit, FIG. 5A is an input image, FIG. 5B is a data string of the input image and the reference image, and FIG. 5C is a power of the input image and the reference image. The spectrum is shown. FIGS. 6A and 6B are diagrams illustrating processing of a feature matching determination unit, FIG. 6A is an input image, FIG. 6B is a data string of the input image and the reference image, and FIG. 6C is a power of the input image and the reference image. The spectrum is shown. It is a figure explaining the process of the two-dimensional temperature data acquisition part of the symbol identification device which concerns on Embodiment 2 of this invention. It is a figure which shows the structure of the symbol identification system containing the symbol identification device which concerns on Embodiment 3 of this invention. FIG. 10 is a diagram illustrating a measurement target region of an infrared camera used in Embodiment 3. It is a figure explaining the centering correction | amendment of a position correction part, and is an example of the centering of an up-down direction. It is a figure explaining the centering correction | amendment of a position correction part, and is an example of the centering of the left-right direction. It is a figure explaining the setting of the determination area | region by the determination area | region setting part. In the symbol identification device concerning Embodiment 5 of this invention, it is a block diagram which shows the internal structure of a two-dimensional temperature data acquisition part. It is a figure which shows the measurement object area | region of the infrared camera used for Embodiment 5. FIG.

Embodiment 1 FIG.
The symbol identification system 1 shown in FIG. 1 measures the temperature distribution on the iron skin surface of the ladle 2 with an infrared camera 4, and the symbol identification device 10 is drawn on the iron skin surface of the ladle 2 from the temperature distribution data. The identification symbol 3 is recognized to identify the ladle 2 individually.

  On the iron skin surface of the ladle 2, the assigned individual identification symbol 3 is drawn with heat-resistant paint. In the example of FIG. 1, a number is drawn as the symbol 3 for individual identification, but any symbol such as a character or a figure may be used.

The infrared camera 4 is installed toward the moving path along which the ladle 2 moves to the next process, and measures a predetermined area including the individual identification symbol 3 of the ladle 2 at the timing when the ladle 2 passes. Then, two-dimensional temperature data representing the temperature distribution in the region is output. Since this infrared camera 4 converts infrared radiation energy into temperature, even if the iron skin and the symbol 3 are actually at the same temperature, the measurement result shows a temperature difference due to the difference in emissivity between the iron skin and the heat-resistant paint. Occurs.
The infrared camera 4 does not need to be prepared exclusively for symbol identification. For example, the infrared camera 4 can also be used as an infrared camera used for leaked steel prediction in Patent Document 1.

  The symbol identification device 10 is configured by a computer and describes the processing contents of the two-dimensional temperature data acquisition unit 11, the filter processing unit 12, the binarization processing unit 13, the reference image holding unit 14, and the feature matching determination unit 15. The symbol identification program is stored in the memory of the computer, and the CPU of the computer executes the symbol identification program stored in the memory.

The two-dimensional temperature data acquisition unit 11 is an input interface of the infrared camera 4, acquires two-dimensional temperature data measured by the infrared camera 4, and has a predetermined data range including a symbol 3 for individual identification (hereinafter referred to as a symbol). Is extracted and output to the filter processing unit 12.
FIG. 2 shows an example of the two-dimensional temperature data 20 and the symbol identification area 21. In this figure, the region where the temperature is high is represented in white, and the region becomes black as the temperature decreases.

The filter processing unit 12 filters the temperature data of the symbol identification region 21 input from the two-dimensional temperature data acquisition unit 11 to remove noise. The noise referred to here is noise derived from the installation environment of the infrared camera 4 (such as reflection of an object existing in the imaging area of the infrared camera 4), noise derived from the shape of the ladle 2 (in the unevenness of the ladle 2) This is noise that occurs in the same manner in the two-dimensional temperature data of the infrared camera 4 every time. Since the appearance location of this noise can be specified in advance, it is removed within a range that does not affect the symbol 3 by filtering.
In addition to removing the inherent noise derived from the installation environment and the shape of the ladle 2, for example, a filtering process such as removing data below a predetermined temperature as noise may be performed.

The binarization processing unit 13 binarizes the temperature data of the symbol identification area 21 from which noise has been removed with a predetermined temperature threshold corresponding to the boundary between the surface temperature of the individual identification symbol 3 and the surface temperature of the iron skin. , Convert to black and white image. For example, for the 32 × 32 pixel symbol identification region 21 in which one temperature measurement value is one pixel, the binarization processing unit 13 first calculates the average temperature of the symbol identification region 21 and uses this as a temporary temperature threshold. . Subsequently, the binarization processing unit 13 lowers the temporary temperature threshold to a low value by ΔT (for example, once), and among the symbol identification regions 21, the temperature region lower than the temporary temperature threshold becomes N pixels. At this point, the temperature threshold is fixed, and a black and white image is generated in which an area having a temperature lower than the temperature threshold is turned on (black) and the remaining area is turned off (white). The temperature threshold at this point corresponds to the boundary between the surface temperature of the symbol 3 for individual identification and the surface temperature of the iron skin, the region of the symbol 3 having a relatively low temperature is black, and the region of the iron skin having a relatively high temperature is The area is converted to white.
Note that N is the same value regardless of the number of the symbol 3 for individual identification, and an appropriate value is set in advance based on the area of the symbol 3 in the symbol identification area 21.

The reference image holding unit 14 is a storage device that stores a reference image used for identifying symbols by the feature matching determination unit 15 in the subsequent stage. Here, the symbol 3 for individual identification of all ladles 2 is measured with the infrared camera 4 in advance, and a black and white image processed by the two-dimensional temperature data acquisition unit 11, the filter processing unit 12 and the binarization processing unit 13 is created. Then, it is stored in the reference image holding unit 14 as a reference image. FIG. 3 shows an example of the reference image of the symbol 3 for individual identification held by the reference image holding unit 14.
Further, the binarization processing unit 13 converts the temperature data in the symbol identification area 21 into a gray scale image and stores it in a storage device such as the reference image holding unit 14 so that the operator can confirm the identification result later. May be.

The feature matching determination unit 15 compares a black and white image (hereinafter referred to as an input image) generated from two-dimensional temperature data obtained by measuring the ladle 2 to be individually identified with each reference image held by the reference image holding unit 14. Then, the reference image having the highest fitness is searched for. The symbol of the reference image is used as the symbol identification result of the input image.
Here, as an example of the determination method, (1) determination based on power spectrum error and (2) determination based on pattern matching will be described.

(1) Determination by Power Spectrum Error FIG. 4 is a diagram illustrating an example of converting two-dimensional temperature data into a one-dimensional data string. The feature matching determination unit 15 converts the input image to be identified by a symbol into a one-dimensional data string as shown in FIG. 4, and performs a fast Fourier transform to calculate a power spectrum. Similarly, the feature matching determination unit 15 converts each reference image held by the reference image holding unit 14 into a one-dimensional data string, and performs a fast Fourier transform to calculate a power spectrum.
Note that the feature matching determination unit 15 may generate a data string in the vertical direction of the image instead of generating a data string in the horizontal direction of the image as shown in FIG. Alternatively, a set of power spectra may be calculated by generating data sequences in the horizontal direction and the vertical direction of the image. Alternatively, the power spectrum may be calculated in two dimensions without converting the two-dimensional temperature data into a one-dimensional data string.
Furthermore, the feature matching determination unit 15 may calculate the power spectrum of the reference image in advance and store it in the reference image holding unit 14.

FIG. 5A shows an input image when the temperature of the ladle 2 provided with the symbol 3 for individual identification “38” is measured. FIG. 5B shows the data string (solid line) of this input image and the data string (broken line) of the reference image of symbol “38”, and FIG. 5C shows the power spectrum of each data string. A reference image of the symbol “38” is shown in FIG.
The feature matching determination unit 15 obtains a square error between the power spectra of the input image and the reference image, and determines that the reference image having a low square error has a high matching degree.
When an input image obtained by capturing the same symbol “38” is compared with a reference image, the waveform pattern does not appear to match in the data string of FIG. However, when compared with the power spectrum of FIG. 5 (c), the square error is as small as about 22.7, and the fitness is high. Thus, the determination based on the power spectrum error can be determined with high accuracy even if there is a positional deviation between ON (black) and OFF (white).

On the other hand, FIG. 6A shows an input image when the temperature of the ladle 2 provided with the symbol 3 for individual identification of “36” is measured. FIG. 6B shows the data string (solid line) of this input image and the data string (broken line) of the reference image with the symbol “38”, and FIG. 6C shows the power spectrum of each data string.
When the input image obtained by imaging the symbol “36” and the reference image obtained by imaging the symbol “38” are compared, the shape of the numbers is similar, and therefore the waveform pattern is similar in the data string of FIG. In comparison with the power spectrum of FIG. 6C, the peak is shifted, the square error is as large as 28.4, and the fitness is low.

(2) Pattern Match The feature match determination unit 15 obtains an area that matches and does not match between the input image and the reference image for each of the ON (black) area and the OFF (white) area. It is determined that a reference image with a high ratio is high in fitness.
When there is little noise in the image, it can be determined with high accuracy by pattern matching.

The feature matching determination unit 15 uses either one of the methods (1) and (2), and determines the symbol of the reference image that has the highest matching degree with the input image (or the matching degree equal to or higher than the standard). The identification result. Alternatively, (1) and (2) may be combined to further improve the determination accuracy. When combining, for example, the feature matching determination unit 15 identifies the symbol of the reference image with the lowest square error of (1) among the reference images in which the pattern match ratio of (2) is equal to or greater than a predetermined value. As a result.
However, even if exception processing is set such that (1) the square error is larger than the upper limit value, it is not applicable, and (2) the pattern match ratio is smaller than the lower limit value. Good.

  As described above, according to the first embodiment, the symbol identification device 10 measures, with the infrared camera 4, the ladle 2 having the individual identification symbol 3 attached to the surface of the iron skin using a heat-resistant paint having a different emissivity. The two-dimensional temperature data acquisition step in which the two-dimensional temperature data acquisition unit 11 acquires the extracted two-dimensional temperature data 20 and extracts the symbol identification region 21, and the filter processing unit 12 removes noise from the temperature data in the symbol identification region 21. Filter processing step, binarization processing unit 13 binarizes the temperature data of the symbol identification area 21 with a predetermined temperature threshold value to generate an input image, and all kinds of symbols for individual identification 3, a reference image measured by the infrared camera 4 and binarized is prepared in the reference image holding unit 14 in advance, and the feature matching determination unit 15 compares each of these reference images with the input image to obtain the most. Symbols if the degree of high reference image and configured to implement a symbol identification method comprising the recognized feature similarity judgment step as a symbol 3 of the input image. For this reason, the infrared camera 4 for temperature management of the ladle 2 can be diverted for individual identification, and a low-cost and high added-value symbol identification method and symbol identification device can be provided.

  Further, according to the first embodiment, the feature matching determination unit 15 obtains the power spectrum of the reference image and the power spectrum of the input image in the feature matching determination step, and the degree of matching is higher as the error between the power spectra is smaller. It was configured to be judged. For this reason, it is possible to make a robust conformance determination against the positional deviation of the symbol 3 for individual identification.

  Further, according to the first embodiment, the binarization processing unit 13 obtains the average temperature of the temperature data of the symbol identification area 21 in the binarization processing step, and lowers the value from the average temperature, and from the value A value at which the number of low temperature data becomes a fixed number (N) is determined as a temperature threshold value and binarized. For this reason, robust conformity determination can be performed with respect to fluctuations in the temperature of the ladle 2.

  Further, according to the first embodiment, the feature matching determination unit 15 is based on the pattern matching between the reference image and the input image in addition to the determination based on the error between the power spectra of the reference image and the input image in the feature matching determination step. The determination is combined to determine the degree of matching. For this reason, the conformity determination can be performed with high accuracy.

Embodiment 2. FIG.
Since the symbol identification system according to the second embodiment is similar to the symbol identification system 1 of FIG. 1 on the drawing, FIG. 1 is used.
It is good to measure with the infrared camera 4 while the ladle 2 is stopped at a predetermined position. However, when measuring while moving to the next process, if the measurement timing is shifted, two-dimensional temperature data is used. Since the position of the symbol 3 for individual identification in 20 is also shifted, the identification accuracy may be lowered. Therefore, in the second embodiment, when the two-dimensional temperature data acquisition unit 11 of the symbol identification device 10 extracts the symbol identification region 21 from the two-dimensional temperature data 20, the extraction position is changed and a plurality of symbol identification regions 21 is obtained. And the feature matching determination unit 15 determines the matching degree of each symbol identification area 21 with the reference image.

FIG. 7 is a diagram showing how nine symbol identification areas 21-1 to 21-9 are extracted from nine locations of the two-dimensional temperature data 20. The two-dimensional temperature data acquisition unit 11 extracts a predetermined data range that is assumed to include the symbol 3 for individual identification from the two-dimensional temperature data 20, and sets it as a symbol identification region 21-1. In addition, a data range of an area moved by a predetermined amount in the upward, downward, right, and left directions is extracted from the symbol identification area 21-1 to be symbol identification areas 21-2 to 21-5. Further, a data range of an area moved by a predetermined amount in the diagonally upper right, diagonally lower right, diagonally upper left, and diagonally lower left directions is extracted from the symbol identification area 21-1, and the symbol identification areas 21-6 to 21-9 are extracted. To do.
Note that FIG. 7 is an example, and the position and number of extraction of the symbol identification region may be arbitrary.

  The filter processing unit 12 performs filter processing on each of the symbol identification areas 21-1 to 21-9 to remove noise. The binarization processing unit 13 determines a temperature threshold value for each of the symbol identification areas 21-1 to 21-9 and generates a binarized black and white image (input image). The feature matching determination unit 15 performs (1) power spectrum error and (2) with each reference image held by the reference image holding unit 14 for each of the nine input images in the symbol identification regions 21-1 to 21-9. ) Determine one or both of the pattern matches. The symbol of the reference image having the highest degree of matching with the nine input images is set as the symbol identification result.

  As described above, according to the second embodiment, the two-dimensional temperature data acquisition unit 11 is assumed to have the symbol 3 for individual identification in the two-dimensional temperature data 20 of the ladle 2 in the two-dimensional temperature data acquisition step. The temperature data of the plurality of symbol identification areas 21-1 to 21-9 are extracted from the areas at different positions, and the binarization processing unit 13 performs a plurality of symbol identification areas 21-1 to 21-1 in the binarization processing step. Each of the temperature data 21-9 is binarized with a predetermined temperature threshold value to generate a plurality of input images, and the feature matching determination unit 15 performs the reference image held by the reference image holding unit 14 in the feature matching determination step. Each is collated with a plurality of input images, and the reference image symbol having the highest degree of matching is recognized. For this reason, even when the infrared camera 4 measures during the movement of the ladle 2 and the positional deviation of the individual identification symbol 3 occurs, the suitability determination can be performed with high accuracy.

Embodiment 3 FIG.
FIG. 8 is a diagram showing a configuration of the symbol identification system 1 according to the third embodiment. The same or equivalent parts as those in FIG. In the third embodiment, the two-dimensional temperature data acquisition unit 11 of the symbol identification device 10 detects the ladle 2 that has moved to the measurement target area of the infrared camera, and acquires the two-dimensional temperature data at an appropriate timing. Realize the configuration. In FIG. 8, only the internal configuration of the two-dimensional temperature data acquisition unit 11 is shown, and the filter processing unit 12, the binarization processing unit 13, the reference image holding unit 14, and the feature matching determination unit 15 are not shown.

  The two-dimensional temperature data acquisition unit 11 according to the third embodiment includes a determination region setting unit 31, an approach detection unit 32, a full / empty pan determination unit (movement direction determination unit) 33, a data acquisition timing determination unit 34, and a two-dimensional temperature. The data storage unit 35 and the position correction unit 36 are included.

  The symbol identification system 1 shown in FIG. 8 detects the ladle 2 that has entered the measurement target area of the infrared camera 4 and measures the temperature distribution on the iron skin surface of the ladle 2. In the illustrated example, the ladle 2 can be moved to the arrows A and B while being lifted by the crane 5. The empty ladle 2 moves in the direction of arrow A, goes to the converter or electric furnace, and receives the refined molten steel. The ladle 2 containing the molten steel moves in the direction of arrow B and goes to the ladle refining facility, and the secondary refining of the molten steel is performed as the next step. The ladle 2 is maintained at a temperature sufficiently higher than the surroundings regardless of the presence or absence of molten steel. Moreover, the moving method of the ladle 2 is not limited to the crane 5, and arbitrary methods may be sufficient.

  The infrared camera 4 is installed toward the moving path of the ladle 2 and outputs two-dimensional temperature data representing the temperature distribution of the region to be measured. FIG. 9 shows a measurement target area 40 of the infrared camera 4. The infrared camera 4 images the measurement target region 40 and converts it into two-dimensional temperature data in which one temperature measurement value forms one pixel. In addition, the 1st determination area | region 41 and the 2nd determination area | region 42 of FIG. 9 are mentioned later.

The determination area setting unit 31 receives the two-dimensional temperature data measured by the infrared camera 4 and, with respect to the two-dimensional temperature data, approaches the ladle 2, the moving direction of the ladle 2, and the presence / absence of molten steel (that is, filling / Empty), and a determination area for determining the data acquisition timing. For example, assuming that the ladle 2 moves in the left-right direction on the measurement target region 40, the determination region setting unit 31 is on the left side of the center of the two-dimensional temperature data measured by the infrared camera 4 as shown in FIG. A first determination area 41 is set, and a second determination area 42 is set on the right side of the center.
The two-dimensional temperature data in which the first determination region 41 and the second determination region 42 are set is output from the determination region setting unit 31 to the approach detection unit 32 and the full / empty pan determination unit 33.

In the approach detection unit 32, in at least one of the first determination region 41 and the second determination region 42, the proportion of pixels equal to or higher than the temperature corresponding to the surface temperature of the ladle 2 (hereinafter, determination temperature T1) is M% or more. And the ladle 2 is determined to have arrived at the measurement target area 40 of the infrared camera 4. M may be any value.
The approach detection result is output from the approach detection unit 32 to the data acquisition timing determination unit 34.

The full / empty pan determination unit 33 determines whether the pixel having the determination temperature T1 or higher appears first from the first determination region 41 or the second determination region 42, and determines the moving direction of the ladle 2 and the presence or absence of molten steel. In this filling / empty pan judgment unit 33, the correspondence between the moving direction of the ladle 2 and filling / empty is set in advance, and in the example of FIG. When the direction is.
When the pixels having the determination temperature T1 or higher appear in the order of the first determination region 41 and the second determination region 42, the ladle 2 is considered to be moving in the direction of the arrow A. It is determined that the ladle 2 is empty. On the other hand, when the pixels having the determination temperature T1 or higher appear in the order of the second determination region 42 and the first determination region 41, it is considered that the ladle 2 is moving in the direction of the arrow B. The part 33 determines that the ladle 2 contains molten steel, that is, is full.
The full / empty determination result is output from the full / empty pan determination unit 33 to the data acquisition timing determination unit 34.

In addition, when the pixel more than determination temperature T1 appears simultaneously in the 1st determination area | region 41 and the 2nd determination area | region 42, the full / empty pan determination part 33 precedes the determination area | region with many pixels with the determination temperature T1 or more first. The charge / empty determination may be performed.
Moreover, in this Embodiment 3, although the case where the ladle 2 moves to the left-right direction of the arrow A and B is assumed, it is also possible to determine a moving direction with respect to the ladle 2 which moves to an up-down direction. is there. This is because the first determination region 41 and the second determination region 42 are shifted in the vertical direction in the measurement target region 40 in advance. When the ladle 2 moves in the vertical direction, the filling / empty pan determination unit 33 causes the ladle 2 to move upward from the bottom when pixels having the determination temperature T1 or higher appear in the order of the first determination region 41 and the second determination region 42. It is determined that it is moving in the direction. On the other hand, when a pixel having the determination temperature T1 or higher appears in the order of the second determination region 42 and the first determination region 41, it is determined that the ladle 2 is moving from top to bottom. In addition, the filling / empty pan determination unit 33 is set to fill / empty according to the up and down movement direction, and the filling / empty pan determination unit 33 performs the filling / empty determination based on the setting. Just do it.

  The data acquisition timing determination unit 34 temporarily holds the two-dimensional temperature data from the time when the approach detection unit 32 is notified of the approach of the ladle 2. In addition, the data acquisition timing determination unit 34 determines the timing at which the pixels having the determination temperature T1 or more become L pixels or more in both the first determination region 41 and the second determination region 42, and the two-dimensional temperature data at this timing is 2 This is stored in the dimension temperature data storage unit 35. L may be set with reference to a value when the ladle 2 is present at the center position of the measurement target region 40 of the infrared camera 4 (that is, two-dimensional temperature data). In addition, according to the determination result of the full / empty pan determination part 33, you may make the value of L variable when the ladle 2 is empty and when it is full.

  However, if the above condition is not reached, the data acquisition timing determination unit 34 times out after the elapse of the predetermined time t1 from the timing when the approach detection unit 32 detects the approach of the ladle 2. In this case, the data acquisition timing determination unit 34 has the largest total number of pixels equal to or higher than the determination temperature T1 of the first determination region 41 and the second determination region 42 among the two-dimensional temperature data temporarily held. The two-dimensional temperature data storage unit 35 stores the two-dimensional temperature data. Further, the fact that the normal acquisition of the two-dimensional temperature data has failed may be output from the data acquisition timing determination unit 34 to the outside.

  In addition, after reaching the above condition, the data acquisition timing determination unit 34 continues for a predetermined time t2 when the pixels having the determination temperature T1 or more become L pixels or less in both the first determination region 41 and the second determination region 42. In this case, it is determined that the ladle 2 has passed the measurement target region 40 of the infrared camera 4, and the determination process is terminated and the temporary holding of the two-dimensional temperature data is also terminated. Further, in this case, an instruction is issued from the data acquisition timing determination unit 34 to the approach detection unit 32 to interrupt the approach detection process for a predetermined time t2, so that the ladle 2 is not erroneously re-detected by the approach detection unit 32. May be.

  In addition, after reaching the above condition, the data acquisition timing determination unit 34 continues for a predetermined time t3 during which at least one of the first determination region 41 and the second determination region 42 has a pixel that is equal to or higher than the determination temperature T1 for L pixels or more. When it does, you may make it judge that the ladle 2 is not moving. In this case, an instruction may be issued from the data acquisition timing determination unit 34 to the approach detection unit 32 to interrupt the next approach detection process of the ladle 2.

  The position correction unit 36 performs centering correction on the two-dimensional temperature data stored in the two-dimensional temperature data storage unit 35, and aligns a data group representing the surface temperature of the ladle 2 with the center of the two-dimensional temperature data. FIG. 10 is a diagram for explaining the vertical centering correction for the two-dimensional temperature data 50, and FIG. 11 is a diagram for explaining the horizontal centering correction. 10 and 11, the two-dimensional temperature data 50 represents a region where the temperature is high in white, and becomes black as the temperature decreases.

(1) Vertical Centering As shown in FIG. 10A, the position correction unit 36 sets a determination slit 51 of a left-right width P pixel at the center of the two-dimensional temperature data 50 and moves upward and downward, respectively. I will let you. During the upward movement, the number of pixels having a temperature lower than the determination temperature T1 × α (for example, α = 0.7 to 1.5) among the P pixels included in the determination slit 51 becomes P / 3. The Y axis is Y1. Similarly, the Y axis where the number of pixels having a temperature lower than the determination temperature T1 × α among the P pixels included in the determination slit 51 during the downward movement is P / 3 is Y2. Finally, the position correction unit 36 performs centering so that the intermediate point between Y1 and Y2 is the center of the two-dimensional temperature data 50, thereby obtaining the two-dimensional temperature data 50a in FIG. 10B is a data area interpolated when the two-dimensional temperature data 50 is centered.

(2) Left / Right Centering As shown in FIG. 11 (a), the position correction unit 36 sets determination slits 52 and 53 having vertical Q pixels at the left end and the right end of the two-dimensional temperature data 50a, respectively, Move to the left. During the movement of the determination slit 52 in the left direction, among the Q pixels included in the determination slit 52, an X axis where the number of pixels having a temperature higher than the determination temperature T1 × α is Q / 3 × 2 or more is denoted by X1. To do. Similarly, during the movement of the determination slit 53 in the right direction, among the Q pixels included in the determination slit 53, the X-axis where the number of pixels having a temperature higher than the determination temperature T1 × α is Q / 3 × 2 or more. , X2. Finally, the position correction unit 36 performs centering so that the intermediate point between X1 and X2 is the center of the two-dimensional temperature data 50a to obtain the two-dimensional temperature data 50b in FIG.
Various numerical values such as P and Q may be arbitrary.

  The two-dimensional temperature data acquisition unit 11 uses a predetermined data range (symbols) including the individual identification symbol 3 from the two-dimensional temperature data 50b after the above processing as described in the first and second embodiments. The identification area 21) is extracted and output to the filter processing unit 12.

  As described above, according to the third embodiment, the symbol identification device 10 has the determination region setting unit 31 for the two-dimensional temperature data input from the infrared camera 4 installed toward the moving path of the ladle 2. A determination region setting step for setting the first determination region 41 and the second determination region 42 at a position corresponding to the movement route, and the data acquisition timing determination unit 34 in both the first determination region 41 and the second determination region 42 are predetermined. A data acquisition timing determination step for determining a timing at which a temperature data group equal to or higher than the determination temperature T1 exceeds a predetermined area (L pixel), and two-dimensional temperature data at the timing determined in the data acquisition timing determination step are stored as two-dimensional temperature data The ladle detection method comprising the data acquisition step stored in the unit 35 is implemented. For this reason, the two determination regions set in the two-dimensional temperature data are monitored, and the ladle 2 is placed at an appropriate position with respect to the infrared camera 4 based on the area of the temperature data group corresponding to the surface temperature of the ladle 2. It is possible to detect the movement, and to acquire and store two-dimensional temperature data at an appropriate timing. Thereby, using the two-dimensional temperature data in which the temperature data group corresponding to the moving ladle 2 appears at the same position every time, individual identification of the ladle 2 is performed, or the possibility of leakage of molten steel is predicted. It becomes possible to do.

  Further, according to the third embodiment, the filling / empty pan determination unit 33 in the order in which the temperature data equal to or higher than the determination temperature T1 appears in the first determination region 41 and the second determination region 42 set in the determination region setting step. Based on this, the moving direction of the ladle 2 is determined, and a moving direction determining step for determining the presence or absence of molten steel according to the moving direction is performed. For this reason, it is possible to determine whether the ladle 2 is empty or not by using the two-dimensional temperature data measured by the infrared camera 4.

  Further, according to the third embodiment, the position correction unit 36 centers the two-dimensional temperature data determined at the data acquisition timing determination step so that the temperature data group equal to or higher than the determination temperature T1 is at the center position. Configured to perform steps. For this reason, even if the timing for acquiring the two-dimensional temperature data is slightly shifted, the position correction is performed, and two-dimensional temperature data in which the temperature data group corresponding to the moving ladle 2 appears at the same position is provided. Can do.

The setting example of the first determination area 41 and the second determination area 42 is not limited to FIG. Hereinafter, another setting example will be described with reference to FIG.
When detecting the ladle 2 that moves in the left-right direction with respect to the measurement target area 40 of the infrared camera 4, the determination area setting unit 31 performs the first determination area 41 and the second determination as shown in FIG. The region 42 is set in a state in which it is arranged side by side. In the case of this setting example, the full / empty pan determination unit 33 determines whether the pixel having the determination temperature T1 or higher appears first from the first determination region 41 or the second determination region 42, and the ladle 2 is left and right. Determine which direction you are moving.

  On the other hand, when detecting the ladle 2 that moves in the vertical direction with respect to the measurement target region 40 of the infrared camera 4, the determination region setting unit 31 includes the first determination region 41 and the first determination region 41 as shown in FIG. 2 The determination areas 42 are arranged in the vertical direction. In the case of this setting example, the full / empty pan determination unit 33 determines whether the pixel having the determination temperature T1 or higher appears first from the first determination region 41 or the second determination region 42, and the ladle 2 is moved up and down. Determine which direction you are moving.

  Alternatively, as shown in FIG. 12C, the first determination area 41 and the second determination area 42 may be arranged vertically and shifted left and right. In this case, similarly to FIG. 9, it is possible to detect the ladle 2 that moves in the vertical direction, the horizontal direction, and the diagonal direction with respect to the measurement target region 40 of the infrared camera 4.

Embodiment 4 FIG.
Since the two-dimensional temperature data acquisition unit 11 shown in the fourth embodiment has the same configuration as that of the third embodiment shown in FIG. 8, the following description will be given with reference to FIG. 8.
When the surface temperature of the ladle 2 is measured using a plurality of infrared cameras 4, the two-dimensional temperature data acquisition unit 11 may perform the processing of Embodiment 3 for each of the infrared cameras 4, or 1 The processing of the third embodiment may be performed only for one infrared camera 4, and the remaining infrared cameras (hereinafter referred to as linked infrared camera 4 a) may perform processing linked to the infrared camera 4. In the fourth embodiment, as an example of linked processing, (1) approach trigger interlocking mode and (2) data acquisition trigger interlocking mode will be described. The infrared camera 4 corresponds to the main camera, and the interlocking infrared camera 4a corresponds to the interlocking camera.

(1) Approach trigger interlocking mode The determination area setting unit 31 receives the two-dimensional temperature data of each of the infrared camera 4 and the interlocking infrared camera 4a, and sets the first determination area 41 and the second determination area 42 for each camera.
The approach detection unit 32 detects the approach of the ladle 2 based on the first determination region 41 and the second determination region 42 set in the two-dimensional temperature data of the infrared camera 4, and the infrared camera 4 and the interlocking infrared camera 4a. The data acquisition timing determination unit 34 is notified as both approach detection results. The data acquisition timing determination unit 34 starts the data acquisition timing determination process for each camera from the time when the approach detection unit 32 is notified of the approach of the ladle 2, and converts the two-dimensional temperature data for each camera into the two-dimensional temperature data. The data is stored in the storage unit 34.

  In the approach trigger interlock mode, since the approach of the ladle 2 to the infrared camera 4 is also regarded as the approach to the interlocking infrared camera 4a, for example, the environment in which the interlocking infrared camera 4a is likely to erroneously detect the approach of the ladle 2 Even if it is installed, it can assist its operation. Therefore, the two-dimensional temperature data acquisition timing determination process following the approach detection can be reliably performed, and the two-dimensional temperature data acquisition failure of the interlocking infrared camera 4a can be prevented.

(2) Data Acquisition Trigger Interlocking Mode The determination area setting unit 31 receives the two-dimensional temperature data of each of the infrared camera 4 and the interlocking infrared camera 4a. Only the two-dimensional temperature data of the infrared camera 4 includes the first determination area 41 and the first determination area 41. 2 The determination area 42 is set.
The approach detection unit 32 detects the approach of the ladle 2 based on the first determination region 41 and the second determination region 42 set in the two-dimensional temperature data of the infrared camera 4 and notifies the data acquisition timing determination unit 34. The data acquisition timing determination unit 34 starts determination processing of the data acquisition timing of the infrared camera 4 from the time when the approach detection unit 32 is notified of the approach of the ladle 2, and at this data acquisition timing, the infrared camera 4 and the interlocking infrared rays are started. Each two-dimensional temperature data of the camera 4a is stored in the two-dimensional temperature data storage unit 34.

  In the data acquisition trigger interlocking mode, the data acquisition timing of the infrared camera 4 is also regarded as the data acquisition timing of the interlocking infrared camera 4a. For example, the interlocking infrared camera 4a easily detects the approach of the ladle 2 or Even if it is installed in an environment where it is easy to misjudge the data acquisition timing, the operation can be assisted. Therefore, the two-dimensional temperature data of the interlocking infrared camera 4a can be acquired more reliably. Further, (1) the processing of the symbol identification device 10 can be further reduced as compared with the approach trigger interlocking mode.

  The full / empty pan determination unit 33 may perform the full / empty determination only for the main infrared camera 4 as a representative of the plurality of cameras. On the other hand, the position correction unit 36 performs centering correction on the two-dimensional temperature data for each camera stored in the two-dimensional temperature data storage unit 34. Thereby, even if the acquisition timing of the two-dimensional temperature data of the interlocking infrared camera 4a slightly deviates from the optimal timing, two-dimensional temperature data in which the data group representing the surface temperature of the ladle 2 is arranged at the center position is obtained. be able to.

  As described above, according to the fourth embodiment, in the approach trigger interlocking mode, the symbol identification device 10 uses one of the plurality of infrared cameras installed toward the moving path of the ladle 2 as the infrared camera 4 and the remaining as the remaining one. As the interlocking infrared camera 4a, the determination area setting unit 31 sets the first determination area 41 and the second determination area 42 at a position corresponding to the movement path of the two-dimensional temperature data input from the infrared camera 4, and A determination region setting step for setting the first determination region 41 and the second determination region 42 at a position corresponding to the movement path of the two-dimensional temperature data input from the interlocking infrared camera 4 a, and the approach detection unit 32 of the infrared camera 4 In at least one of the first determination area 41 and the second determination area 42 set for the two-dimensional temperature data, (M%) of temperature data equal to or higher than a predetermined determination temperature T1. When the appearance is detected and the approach detection step determines the approach of the ladle 2 to the infrared camera 4 and the approach detection step determines the approach, the data acquisition timing determination unit 34 determines whether the infrared camera 4 is two-dimensional. In both the first determination area 41 and the second determination area 42 set for the temperature data, the timing at which the temperature data group equal to or higher than the determination temperature T1 becomes equal to or greater than a predetermined area (L pixel) is determined, and the interlocking infrared camera Data acquisition for determining the timing at which the temperature data group equal to or higher than the determination temperature T1 exceeds the predetermined area (L pixel) in both the first determination area 41 and the second determination area 42 set for the two-dimensional temperature data of 4a. Two-dimensional temperature data of the infrared camera 4 at each timing determined in the timing determination step and the data acquisition timing determination step And a data acquisition step of storing the 2-dimensional temperature data of the pre-interlocking infrared camera 4a in each two-dimensional temperature data storage unit 34 is configured to implement. For this reason, it becomes possible to assist the interlocking infrared camera 4a installed in an environment in which the approach of the ladle 2 is likely to be erroneously detected, and not only the infrared camera 4 but also the two-dimensional temperature data of the interlocking infrared camera 4a. It can be acquired and stored at an appropriate timing.

  Further, according to the fourth embodiment, in the data acquisition trigger interlocking mode, the symbol identification device 10 sets one of the plurality of infrared cameras installed toward the moving path of the ladle 2 as the infrared camera 4 and the remaining as the remaining one. The determination infrared region 4a is set as the interlocking infrared camera 4a, and the determination region setting unit 31 sets the first determination region 41 and the second determination region 42 at a position corresponding to the movement path of the two-dimensional temperature data input from the infrared camera 4. In both the first determination area 41 and the second determination area 42 set by the step and the data acquisition timing determination unit 34 for the two-dimensional temperature data of the infrared camera 4, a temperature data group equal to or higher than the determination temperature T <b> 1 has a predetermined area ( L pixel) data acquisition timing determination step for determining the above timing, and infrared at the timing determined by the data acquisition timing And configured to perform a data obtaining step of storing camera 4 of the two-dimensional temperature data and the two-dimensional temperature data interlocking infrared camera 4a in each two-dimensional temperature data storage unit 34. For this reason, it becomes possible to assist the interlocking infrared camera 4a installed in an environment in which erroneous detection of the approach of the ladle 2 and data acquisition timing can be easily determined, and not only the infrared camera 4 but also the interlocking infrared camera. The two-dimensional temperature data 4a can also be acquired and stored at an appropriate timing.

Embodiment 5 FIG.
FIG. 13 is a block diagram illustrating an internal configuration of the two-dimensional temperature data acquisition unit 11 in the symbol identification device 10 according to the fifth embodiment. 13 that are the same as or equivalent to those in FIG. 8 are given the same reference numerals, and descriptions thereof are omitted. FIG. 14 shows a measurement target region 40 of the infrared camera 4 used by the symbol identification device 10 according to the fifth embodiment. In the fifth embodiment, it is assumed that the ladle 2 moves in the vertical direction of the arrow C as shown in FIG. 14, and further, a foreign object other than the ladle 2 (detection target) on the movement path of the ladle 2. Assume that the moving routes of other ladles etc. overlap partially. As an example, the movement path of the foreign object is the left-right direction of the arrow D.

  In the symbol identification device 10 illustrated in FIG. 13, the determination region setting unit 31 a is a first determination region serving as a determination region for determining the approach of the ladle 2, the full / empty determination, and the data acquisition timing as illustrated in FIG. 14. 41 and the second determination area 42 are set in the measurement target area 40 of the infrared camera 4. In addition, the determination area setting unit 31a sets the foreign object determination areas 43 and 44, which are determination areas for detecting a foreign object, in the measurement target area 40. The foreign matter determination areas 43 and 44 may be set at least one place on the assumed foreign matter movement path.

  When at least one of the foreign substance determination areas 43 and 44 has a pixel equal to or higher than the temperature corresponding to the foreign substance (hereinafter referred to as a determination temperature T2), the foreign object detection unit 37 detects the foreign object as the measurement target area of the infrared camera 4. 40 is determined to be moving, and is notified to the data acquisition timing determination unit 34a. R may be any value.

  The data acquisition timing determination unit 34a starts the determination of the data acquisition timing from the time when the approach detection unit 32 is notified of the approach of the ladle 2, but when the foreign object detection is notified from the foreign object detection unit 37, the predetermined time t4. Only interrupt this decision. Therefore, the two-dimensional temperature data storage unit 35 does not store the two-dimensional temperature data for a predetermined time t4. Thereby, it is possible to prevent the two-dimensional temperature data measured while foreign substances other than the ladle 2 are in the first determination area 41 and the second determination area 42 from being erroneously acquired and stored.

  As described above, according to the fifth embodiment, the determination region setting unit 31a has the foreign object of the two-dimensional temperature data output from the infrared camera 4 when the movement path of the ladle 2 and the movement path of the foreign object partially overlap. The foreign matter determination areas 43 and 44 are set at positions corresponding to the movement path of the foreign matter, and the foreign matter detection unit 37 has a temperature data group equal to or higher than a predetermined determination temperature T2 in the foreign matter determination areas 43 and 44 to a predetermined area (R pixel) or more. A foreign matter detection step for determining a foreign matter when the foreign matter is detected, and the data acquisition timing determination unit 34a is configured to interrupt the data acquisition timing determination process when a foreign matter is detected in the foreign matter detection step. For this reason, it is possible to prevent the two-dimensional temperature data storage unit 34 from acquiring and storing the two-dimensional temperature data obtained by measuring the surface temperature of the foreign object by mistake with the ladle 2.

Although the embodiment of the present invention has been described in detail with reference to the drawings, the specific configuration is not limited to the configuration of the above-described embodiment, and the design does not depart from the gist of the present invention. Needless to say, changes and the like are included in the present invention.
Moreover, although the case where the symbol identification method of this invention is used for the individual identification of a ladle was demonstrated to the example, it is not limited to this, It can apply also to objects other than a ladle.

DESCRIPTION OF SYMBOLS 1 Symbol identification system 2 Ladle 3 Symbol for individual identification 4 Infrared camera 5 Crane 10 Symbol identification device 11 Two-dimensional temperature data acquisition part 12 Filter processing part 13 Binarization processing part 14 Reference | standard image holding part 15 Feature conformity judgment part 20 Two-dimensional temperature data 21, 21-1 to 21-9 Symbol identification region 31 Determination region setting unit 32 Approach detection unit 33 Filling / empty pan determination unit 34 Data acquisition timing determination unit 35 Two-dimensional temperature data storage unit 36 Position correction unit 37 Foreign object detector 40 Measurement target area 41 First determination area 42 Second determination area 43, 44 Foreign object determination area 50 Two-dimensional temperature data 51-53 Determination slit

Claims (12)

A two-dimensional temperature data acquisition step of acquiring two-dimensional temperature data obtained by measuring an object having an individual identification symbol attached to the surface using materials having different emissivities with an infrared camera;
A binarization processing step of binarizing the two-dimensional temperature data acquired in the two-dimensional temperature data acquisition step with a predetermined temperature threshold to generate an input image;
For all the types of symbols for individual identification, reference images measured with the infrared camera and binarized are prepared in advance, and each of the reference images is collated with the input image generated in the binarization processing step. And a feature matching determination step of recognizing the symbol of the reference image having the highest matching level as the symbol of the input image.   2. The symbol according to claim 1, wherein in the feature matching determination step, a power spectrum of the reference image and a power spectrum of the input image are obtained, and it is determined that the matching degree is higher as an error between the power spectra is smaller. Identification method. In the two-dimensional temperature data acquisition step, a plurality of two-dimensional temperature data are extracted at different positions from a region where the symbol for individual identification is assumed to exist in the two-dimensional temperature data obtained by measuring the object,
In the binarization processing step, each of the plurality of two-dimensional temperature data is binarized with a predetermined temperature threshold to generate a plurality of input images,
The feature matching determination step is characterized by recognizing the reference image symbol having the highest degree of matching by comparing each of the reference images of all types of the individual identification symbols with the plurality of input images. The symbol identification method according to claim 1 or claim 2.   In the binarization processing step, an average temperature of the two-dimensional temperature data is obtained, and a value that lowers the value from the average temperature and sets the number of data at a temperature lower than that value to a constant number is defined as a binary value. 4. The symbol identification method according to claim 1, wherein:   In the feature adaptation determination step, in addition to the determination based on an error between power spectra of the reference image and the input image, a determination based on a pattern match between the reference image and the input image is combined to determine the degree of matching. The symbol identification method according to any one of claims 1 to 4, characterized in that: The two-dimensional temperature data acquisition step includes
A determination area setting step of setting two determination areas at a position corresponding to the movement path of the two-dimensional temperature data output by the infrared camera installed toward the movement path of the object;
Data acquisition for determining a timing at which a temperature data group equal to or higher than a predetermined determination temperature is equal to or greater than a predetermined area in both of the two determination regions set in the determination region setting step, and acquiring the two-dimensional temperature data at the timing The symbol identification method according to claim 1, further comprising a timing determination step. The two-dimensional temperature data acquisition step includes
One of the plurality of infrared cameras installed toward the moving path of the object is a main camera, and the remaining is an interlocking camera, which corresponds to the moving path of the two-dimensional temperature data output by the main camera. A determination area setting step for setting two determination areas at positions, and setting two determination areas at positions corresponding to the movement path of the two-dimensional temperature data output by the interlocking camera;
By detecting the appearance of temperature data equal to or higher than a predetermined determination temperature in at least one of the two determination regions set for the two-dimensional temperature data of the main camera, the approach of the object to the main camera is detected. An approach detection step for determining;
When approach is determined in the approach detection step, a temperature data group equal to or higher than the predetermined determination temperature is equal to or greater than a predetermined area in both of the two determination regions set for the two-dimensional temperature data of the main camera. The timing is determined to acquire the two-dimensional temperature data at the timing, and the temperature data equal to or higher than the predetermined determination temperature in both of the two determination areas set for the two-dimensional temperature data of the interlocking camera. 6. The data acquisition timing determination step of determining a timing at which the group becomes a predetermined area or more and acquiring the two-dimensional temperature data at the timing. 6. The symbol identification method described. The two-dimensional temperature data acquisition step includes
One of the plurality of infrared cameras installed toward the moving path of the object is a main camera, and the remaining is an interlocking camera, which corresponds to the moving path of the two-dimensional temperature data output by the main camera. A determination area setting step for setting two determination areas at positions;
In both of the two determination areas set for the two-dimensional temperature data of the main camera, a timing at which a temperature data group equal to or higher than the predetermined determination temperature becomes a predetermined area or more is determined, and the two-dimensional temperature at the timing is determined. The symbol identification method according to any one of claims 1 to 5, further comprising a data acquisition timing determination step of acquiring data. When a movement path of the object and a movement path of a foreign object other than the object partially overlap, a foreign object determination area is set at a position corresponding to the movement path of the foreign object in the two-dimensional temperature data output by the infrared camera. A foreign object detection step for determining a foreign object when a temperature data group equal to or higher than a predetermined foreign object determination temperature is equal to or greater than a predetermined area in the foreign object determination region;
The symbol according to any one of claims 6 to 8, wherein, in the data acquisition timing determination step, the data acquisition timing determination process is interrupted when a foreign object is detected in the foreign object detection step. Identification method.   7. A moving direction determination step for determining a moving direction of the object based on the order in which temperature data equal to or higher than the predetermined determination temperature appears in the two determination areas set in the determination area setting step. The symbol identification method according to claim 9.   7. A centering correction step of centering the two-dimensional temperature data at the timing determined in the data acquisition timing determination step so that a temperature data group equal to or higher than the predetermined determination temperature becomes a center position. The symbol identification method according to claim 10. A two-dimensional temperature data acquisition unit for acquiring two-dimensional temperature data obtained by measuring an object having an individual identification symbol attached to a surface using materials having different emissivities with an infrared camera;
A binarization processing unit that binarizes the two-dimensional temperature data acquired by the two-dimensional temperature data acquisition unit with a predetermined temperature threshold, and generates an input image;
For all types of symbols for individual identification, a reference image holding unit holding a reference image measured and binarized by the infrared camera,
Feature matching in which each reference image held by the reference image holding unit is compared with the input image generated by the binarization processing unit, and the symbol of the reference image having the highest degree of matching is recognized as the symbol of the input image A symbol identification device comprising a determination unit.

Images